Thanks for being a day lead for NeuroAI! We're so glad you're accepted to help. We have an opportunity to help define a burgeoning field! Below are some instructions specific to the course–we'll go over this information in small group settings as well, but you can always use this for reference.
The overall course outline is available here. Have a look at the course philosophy, the learner profiles and the learning objectives for your day. You can also dig deeper into related days to help better understand where students will be at when they take your day.
Making a one-pager for a day
Your first task is to fill in the "one-pager" for a day. Make a copy of the day outline template for your day. The coarse-grained learning objectives can be copied from the course outline. As day lead, you have a lot of flexibility in how you present your materials as long as it fits within the learning objectives for the day. The curriculum team will have a look at the one-pager once it's filled in and iterate with you.
The one-pager template is organized such that the tutorial outlines are listed first. In our experience, it's better to focus on the tutorials first, as they take the most time to get right. The intro lecture can then serve to supplement and give a larger context for the tutorial; be aware, however, that while all students will go through the tutorials, some may opt to skip the intro lecture.
The computational neuroscience (CN) and deep learning (DL) courses are prerequisites for the NeuroAI course. We expect the students to 1) be very familiar with the format of the lectures and exercises and 2) remember some of the concepts covered therein (though it never hurts to review). Therefore, we recommend:
.1If you've never made or experienced an NMA class before, take some time to go through an entire day or at least a whole tutorial, at your leisure. This will help you familiarize yourself with the house style and the common types of materials and exercises we use.
.2Search through DL and CN courses for relevant references and related concepts. Linking your class to previous ones can help you tune your materials to students' existing knowledge. You might also be able to re-use some assets, e.g. data loaders, models. As part of your one-pager, you should list pre-requisites.
We use PyTorch exclusively in the class, not jax, as PyTorch was used by the DL class. Some of the other libraries we use include:
numpy
scipy
matplotlib & seaborn
pandas (use sparingly)
gradio
If you want to add unusual libraries to the mix, let us know! We aim to keep the total number of libraries can make things simpler for students and building the final book, but can accomodate when there are common needs.
Notation
We will add a notation guide later on during course prep